基于单目视觉的矿井机车障碍物检测和测距方法

Obstacle detection and ranging method of mine locomotive based on monocular vision

针对现有矿井机车障碍物检测和测距方法存在测距精度低、测量范围小及成本高等问题,提出了一种基于单目视觉的矿井机车障碍物检测与测距方法。该方法首先对CCD摄像机采集的图像进行预处理,然后根据障碍物的特点进行特征提取得到准确的障碍物区域,最后利用基于摄像机内部参数和几何关系的单目视觉测距方法得到机车与前方障碍物之间的距离。测试结果表明,该方法能有效检测机车前方障碍物,且测距精度在有效误差范围内。

In view of problems of low ranging accuracy,small measuring range and high cost of existing obstacle detection and ranging method of mine locomotive,an obstacle detection and ranging method of mine locomotive based on monocular vision was proposed.Firstly,the method preprocesses image captured by CCD camera,then extracts feature according to characteristics of the obstacle,so as to get accurate obstacle region.Finally,it uses monocular vision ranging method based on camera interior parameters and geometrical relationship to get distance between locomotive and the preceding obstacle.The test results show that the method can detect the obstacle ahead of locomotive effectively and ranging accuracy is within effective error range.